Conventionally, there is known a method of performing a link operation in a hyper-media environment using videos, where an anchor is displayed on a subject of the video, a link is tied to the anchor, and the anchor is specified in the video. In the following, the subject is also referred to as an object.
For example, there is a technology of setting an anchor in an arbitrary area moving in the video, storing anchor information and link information, and dynamically displaying the anchor based on display attributes of the video and the anchor.
Patent document 1: Japanese Laid-Open Patent Publication No. H8-123816
An anchor may be set to an object in the video by a method of, for example, extracting the shape of the object by extracting edges in each frame in the video, and furthermore, extracting the error between frames, to extract the movement of the object. An anchor is, for example, an element set in a hyperlink.
When an object that is the target for setting an anchor moves in the video, in the conventional technology, the position coordinates of the object in each frame are identified, and the position of the anchor is defined by the frame number and the position coordinates of the object.
Meanwhile, there may be cases of editing the anchor that has already been set in the object. For example, it is assumed that an object is extracted by edge extraction, and an anchor positioned at this object. Depending on the precision of edge extraction, there may be a difference between the position or shape of the object in an image viewed by a person and the position or shape of the object recognized by the computer according to edge extraction. In this case, there arises the operation of changing the position of the anchor in the frame so that the anchor is superposed on the object as viewed by a person. In the following, this operation is also referred to as an editing operation.
However, if the position coordinates of the object in each frame are identified and the position of the anchor is defined by a frame number and position coordinates, there will be a number of definitions of anchor positions corresponding to the number of frames. Therefore, the data amount relevant to anchor positions becomes large.
Assuming that an operation of editing the anchor position is performed for the definitions of anchor positions present in the respective frames, changes will be made for the respective definition data items corresponding to the number of frames. This is a time-consuming process for the editor.